Akamai state of_the_internet_q2_2010
-
Upload
constantin-cocioaba -
Category
Documents
-
view
649 -
download
0
description
Transcript of Akamai state of_the_internet_q2_2010
2nd Quarter, 2010 Report
Volume 3, Number 2
The State of the Internet
Get the most out of the State of the Internet
www.akamai.com/stateoftheinternet
Understand what’s happening now
with real-time data visualizations
Learn abut historical trends through
archived reports and webcasts
Connect with us on Twitter
Executive SummaryEach quarter, Akamai publishes a “State of the Internet” report. This report includes data gathered from across Akamai’s global server network about attack traffic, broadband adoption, and mobile connectivity, as well as trends seen in this data over time. Periodically, it also aggregates publicly available news and information about notable events seen throughout the quarter, including Denial of Service attacks, Web site hacks, and network events, including outages and new connections.
During the second quarter of 2010, Akamai observed attack traffic originating from 200 unique countries/regions
around the world. The United States was the top attack traffic source, accounting for 11% of observed attack
traffic in total. China and Russia held the second and third place spots respectively, accounting for just over
20% of observed attack traffic. Attack traffic concentration was lower in the second quarter, with the top
10 ports seeing just over 86% of the observed attack traffic. Aggregated at a continental level, Europe continued
to be responsible for the highest percentage of attacks seen in the second quarter. Port 445 continued to be the
most highly targeted port for observed attacks, though attacks targeted at Port 23 (Telnet) and Port 22 (SSH) both
grew during the second quarter.
Akamai observed a 2.8% increase (from the first quarter of 2010) globally in the number of unique IP addresses
connecting to Akamai’s network, pushing past the 500 million mark for the first time. From a global connection
speed perspective, South Korea continued to have the highest level of “high broadband” (>5 Mbps) connectivity,
with 3 of every 4 connections to Akamai at speeds above 5 Mbps. South Korea also maintained the highest
average connection speed, at 17 Mbps (up from 12 Mbps in the first quarter), and recorded the highest average
peak6 connection speed, at 38 Mbps, where the per-IP address peak connection speed was averaged across the
IP addresses seen from each country. Cities in South Korea continued to hold many of the top spots in the rankings
of highest average and average peak connection speeds by city as well. In the United States, Delaware remained
in the top position, with 67% of connections to Akamai occurring at 5 Mbps or greater. Delaware also maintained
the highest average connection speed in the United States, though it declined to 7.2 Mbps, and once again re-
corded the highest average peak connection speed across the United States, at 24 Mbps. Monterey Park, CA was
the United States city with the highest average connection speed (6.9 Mbps) and highest average peak connection
speed (26 Mbps).
Second quarter attack traffic from known mobile networks was more concentrated than overall observed attack
traffic, with half of the observed mobile attacks coming from just the top three countries. Ports targeted by mobile
attack traffic were similar to the overall port list, though a higher percentage of attacks targeting Port 445 were
observed when looking at just known mobile networks. In the second quarter of 2010, average measured
connection speeds on mobile network providers around the world ranged from 6.1 Mbps down to 115 Kbps
– once again, both were observed on mobile providers in Slovakia. Of the 109 mobile network providers listed
in the report, 95 saw average connection speeds increase in the second quarter. Average peak connection speeds
on mobile providers around the world ranged from over 20 Mbps down to just over 700 Kbps. In reviewing content
consumption metrics, we found that the amount of content downloaded from Akamai per month per unique
IP address showed quarterly growth in 71 of 109 providers, and yearly growth in 85 of those providers.
4 © 2010 Akamai Technologies, Inc. All Rights Reserved
IntroductIon 5
SectIon2:SecurIty 6
2.1 Attack Traffic, Top Originating Countries 6
2.2 Attack Traffic, Top Ports 7
SectIon3:InternetPenetratIon 8
SectIon4:GeoGraPhy– Global 9
4.1 Global Average Connection Speeds 10
4.2 Global Average Connection Speeds, City View 11
4.3 Global Average Peak Connection Speeds 13
4.4 Global Average Peak Connection Speeds, City View 14
4.5 Global High Broadband Connectivity 16
4.6 Global Broadband Connectivity 17
4.7 Global Narrowband Connectivity 18
SectIon5:GeoGraPhy–unItedStateS 19
5.1 United States Average Connection Speeds 19
5.2 United States Average Connection Speeds, City View 20
5.3 United States Average Peak Connection Speeds 20
5.4 United States Average Peak Connection Speeds, City View 21
5.5 United States High Broadband Connectivity 22
5.6 United States Broadband Connectivity 23
5.7 United States Narrowband Connectivity 24
SectIon6:MobIle 25
6.1 Attack Traffic, Top Originating Countries 25
6.2 Attack Traffic, Top Ports 26
6.3 Mobile Connection Speeds & Data Consumption 27
SectIon7:aPPendIx 30
SectIon8:endnoteS 31
Table of Contents
5© 2010 Akamai Technologies, Inc. All Rights Reserved
IntroductionAkamai’s globally distributed network of servers allows us to gather massive amounts of information on many metrics, including connection speeds, attack traffic, and network connectivity/availability/latency problems, as well as traffic patterns on leading Web sites.
In the second quarter of 2010, observed attack traffic continued to target a relatively consistent set of ports,
and attacks targeting port 445 continued to be responsible for a significant percentage of the observed attacks,
both overall, and for attacks observed to be originating in known mobile networks. The United States was the top
source of observed attack traffic, followed closely by China – in aggregate, the two countries were responsible
for 22% of observed attack traffic, with the balance originating in a long tail of 198 additional countries/regions.
Aggregated at a continental level, Europe was responsible for the highest percentage of observed attacks.
Quarterly growth in the global observed unique IP count continued, up almost 3% from the prior quarter,
with over 500 million unique IP addresses making requests to Akamai in the second quarter. Quarterly trending
in global average connection speeds was mixed during the second quarter, though quarterly trending in average
peak6 connection speeds was generally positive. The overall global trends for high broadband adoption were
generally positive as well, with 1.6x as many countries showing quarterly gains as losses, and 1.7x as many countries
showing yearly gains as losses. In looking at broadband adoption levels in the second quarter, 55 countries/regions
had broadband adoption levels in excess of 50% -- up from 50 in the prior quarter, and 41 a year earlier. Rates
of narrowband adoption were generally lower around the world, with less than five percent of connections
to Akamai overall at speeds below 256 Kbps in the second quarter.
In the United States, Delaware continued to maintain its position as the state with the strongest connection speed
metrics, placing first for average connection speed (7.2 Mbps) and average peak connection speed (24 Mbps),
as well as high broadband (67%) and broadband (98%) adoption.
In response to the growing amount of Internet content being accessed through mobile devices such as smartphones
and laptops equipped with mobile broadband connection technologies, and also in response to multiple inquiries
for such data, Akamai continues to publish insights into metrics collected from connections to Akamai that have
been identified as coming from networks associated with mobile providers. Second quarter attack traffic from
known mobile networks was more concentrated than overall observed attack traffic, with half of the observed
mobile attacks coming from just the top three countries. Ports targeted by mobile attack traffic were similar
to the overall port list, though a higher percentage of attacks targeting Port 445 were observed when looking
at just known mobile networks. During the second quarter of 2010, average measured connection speeds on
mobile network providers around the world ranged from 6.1 Mbps, down to 115 Kbps – both were observed
on mobile providers in Slovakia. Average peak connection speeds on mobile providers around the world ranged
from over 20 Mbps down to just over 700 Kbps. Consumption of content from Akamai ranged from close
to 15 GB per unique IP per month down to just 14 MB per unique IP per month, nearly a 1000x difference.
6 © 2010 Akamai Technologies, Inc. All Rights Reserved
2.1 Attack Traffic, Top Originating CountriesDuring the second quarter of 2010, Akamai observed
attack traffic originating from 200 unique countries/
regions, up two from the prior quarter. After three
consecutive quarters topping the list, Russia fell to third
place, generating only 10% of observed attack traffic
in the second quarter. As shown in Figure 1, the United
States and China moved up to top the list, with both
countries originating 11% of observed attack traffic
in the second quarter. Taiwan, Brazil, Italy, Germany,
Romania, and Japan held steady from quarter-to-
quarter, maintaining their placement on the top 10
list, while Turkey replaces Poland in 10th place. Attack
concentration among the top 10 countries continued
to decline, dropping to 58% in the second quarter.
Once again aggregating the observed attack traffic at a
continental level, we find that the percentage of attacks
that Europe was responsible for dropped approximately
11% in the second quarter to 39%, while all of the other
geographies saw corresponding increases. Aggregating this
data over the last two years, as shown in Figure 2, illustrates
some interesting trends. Except for Africa, all of the conti-
nents within the graph saw their most significant changes
occur in the Q2-Q3 2009 time period. This may be related
to the rise of the Conficker family of worms – their growth
in activity was covered in depth in 2009 editions of the
State of the Internet report series. Africa’s consistently low
contribution to the overall set of observed attacks is of little
surprise, given the generally poor state of Internet connec-
tivity in most countries on the African continent. (However,
it is worth noting that observed attack traffic from Africa
more than doubled to 2% from the first quarter to the
second quarter of 2010.) During the Q2-Q3 2009 “inflection
quarter,” Europe’s contribution nearly doubled, while the
percentage of attacks observed from the Asia Pacific region
declined by nearly 40%. Similar changes were seen within
the Americas, where South America’s percentage more
than doubled during that quarter, while North America’s
contribution dropped by half.
SECTION 2:
Security
Akamai maintains a distributed set of agents deployed across the Internet that serve to monitor attack traffic. Based on the data collected by these agents, Akamai is able to identify the top countries from which attack traffic originates, as well as the top ports targeted by these attacks. (Ports are network layer protocol identifiers.) This section provides insight into Internet attack traffic, as observed and measured by Akamai, during the second quarter of 2010.
Figure 1: Attack Traffic, Top Originating Countries/Regions
1 United States 11% 10%
2 China 11% 9.1%
3 Russia 10% 12.0%
4 Taiwan 6.0% 6.1%
5 Brazil 6.0% 6.0%
6 Italy 3.0% 4.4%
7 Germany 3.0% 3.9%
8 Romania 3.0% 3.2%
9 Japan 3.0% 2.9%
10 Turkey 3.0% 1.5%
– Other 42% 39%
Q1 ‘10 %% TrafficCountry/Region
41
87 10
3
5
2
6
9
7© 2010 Akamai Technologies, Inc. All Rights Reserved
2.2 Attack Traffic, Top PortsAttack traffic concentration among the top 10 targeted
ports declined significantly quarter-over-quarter, with the
top 10 ports responsible for just under 87% of observed
attacks (down from just under 95% in the first quarter of
2010). Port 445 continued to be the most targeted port,
though attack traffic to it declined approximately 12%
from the first quarter of 2010. We will continue to watch
this metric going forward to determine if attacks targeted
at Port 445 are truly declining, or if this was simply a one
quarter anomaly. The growth in attacks targeted at Port
23 (Telnet) may be related to similar observations1 made by
others, who noted an increase in unsolicited connections
to that port that completed the three-way handshake and
then were closed. While traffic targeted at Port 22 (SSH)
declined slightly from the first quarter from a percentage
perspective, the number of actual observed attacks actually
increased. Similar to the growth in Telnet-targeted attacks
noted above, this increase is in line with an increase noted2
by others, who reported that SSH brute-force login attempts
appeared to be on the rise again in the second quarter.
In the second quarter, Port 1080 (SOCKS Proxy) was forced
from the top 10 list, replaced by newcomer Port 3389 (Mi-
crosoft Terminal Services). In looking at the port distribution
among the top 10 countries, it is interesting to note that
Port 23 was among the top five targeted ports in Brazil,
Italy, Russia, and Turkey – in Turkey, it was even responsible
for nearly 2.5x the attacks targeted at Port 445. That port
continued to be, far and away, the most targeted port in
the other countries among the top 10, responsible for any-
where from 1.1x to 64x the number of attacks as the next
most targeted port (in China and Romania respectively.)
Figure 3: Attack Traffic, Top Ports
445 Microsoft-DS 62% 74%
23 Telnet 9.2% 2.5%
22 SSH 6.1% 6.3%
135 Microsoft-RPC 3.0% 2.5%
139 NetBIOS 1.9% 3.2%
1433 Microsoft SQL Server 1.2% 1.1%
4899 Remote Administrator 1.2% 1.5%
80 WWW 1.1% 1.7%
5900 VNC Server 0.7% 0.9%
3389 Microsoft Terminal Services 0.4% 0.3%
Various Other 13.6% –
Q1 ‘10 %% TrafficPort UsePort VNC Server 0.7%Microsoft Terminal Services 0.4%
Microsoft-RPC 3.0%NetBIOS 1.9%
Microsoft SQL Server 1.2%Remote Administrator 1.2% WWW 1.1%
SSH 6.3%
Other 5.6%
Microsoft-DS 62%
Other13.6%SSH
6.1%
Telnet9.2%
Figure 2: Historical Observed Attack Traffic, Aggregated by Region
Africa
South America
North America
Asia Pacific
Europe
60%
% o
f O
bser
ved
Att
acks
50%
40%
30%
20%
10%
0%
Q3 08 Q3 09Q2 08 Q2 09Q4 08 Q4 09Q1 09 Q2 10Q1 10
8 © 2010 Akamai Technologies, Inc. All Rights Reserved
As shown in Figure 4, the top 10 countries remained the
same quarter-over-quarter, though Brazil and Canada
exchanged places, with Brazil slowly climbing the list over
the past several quarters. While the last several State of
the Internet reports have noted that the United States
and Canada have accounted for approximately 40% of
the observed IP addresses, it appears that growth in these
two countries may be slowing, relative to other countries
around the world, as they accounted for a combined
38% of the observed IP addresses in the second quarter.
Two of the top ten countries (the United Kingdom and
Canada) showed quarterly declines, though each were
less than one percent, while the remaining eight coun-
tries showed modest quarterly increases.
Globally, 59 countries/regions saw a quarterly decline in IP
address counts in the second quarter, although just over a
third of those were places where Akamai observed just tens
or hundreds of IP addresses. Concentration among the top
ten remained fairly consistent as well, accounting for ap-
proximately 70% of the observed IP addresses. In looking at
the “long tail,” there were 186 countries/regions with fewer
than one million unique IP addresses connecting to Akamai
in the second quarter of 2010, 139 with fewer than 100,000
unique IP addresses, and 34 with fewer than 1,000 unique
IP addresses. The country counts for the one million and
one thousand unique IP address thresholds were up slightly
quarter-over-quarter, while the 100,000 address threshold
lost only a single country.
SECTION 3:
Internet Penetration
Through a globally-deployed server network, and by virtue of the billions of requests for Web content that it services on a daily basis, Akamai has unique visibility into the levels of Internet penetration around the world. In the second quarter of 2010, over 500 million unique IP addresses, from 236 countries/regions, connected to the Akamai network – 2.8% more IP addresses than in the first quarter of 2010, and 18% more than in the same quarter a year ago. Although we see approximately 500 million unique IP addresses, Akamai believes that it sees well over one billion Web users. This is because, in some cases, multiple individuals may be represented by a single IP address (or small number of IP addresses), as they access the World Wide Web through a firewall or proxy server. Conversely, individual users can have multiple unique IP addresses associated with them, due to their use of handheld devices, personal/home computers, business computers, Internet-connected video/gaming systems, etc.
Figure 4: Unique IP Addresses Seen By Akamai
– Global 501,047,374 2.8% 18%
1 United States 131,397,342 1.6% 15%
2 China 60,042,886 4.0% 30%
3 Japan 34,833,550 4.9% 14%
4 Germany 31,220,592 0.7% 6.9%
5 France 22,740,013 1.2% 13%
6 United Kingdom 20,042,782 -0.4% 6.6%
7 South Korea 16,795,947 0.5% 15%
8 Brazil 12,003,674 5.5% 19%
9 Canada 11,644,413 -0.7% 8.1%
10 Spain 11,444,687 2.0% 13%
YoY Change
QoQ Change
Q2 ‘10 Unique IP Addresses
Country/Region
1
9
4
8
23
710
5
6
9© 2010 Akamai Technologies, Inc. All Rights Reserved
The data presented within this section was collected
during the second quarter of 2010 through Akamai’s
globally-deployed server network and includes all
countries/regions that had more than 1,000 unique
IP addresses make requests to Akamai’s network during
the second quarter. For the purposes of classification in
this report, the “broadband” data included below is for
connections greater than 2 Mbps, and “high broadband”
is for connections of 5 Mbps or greater. (In July 2010,
the United States Federal Communications Commission
(FCC) revised its working definition of broadband from
services in “excess of 200 kilobits per second (kbps)
in both directions” – a standard adopted over a decade
ago in the 1999 First Broadband Deployment Report
– to “service offering actual download (i.e., to the
customer) speeds of at least 4 Mbps”.4 While we have
not adopted this revised definition of broadband for
this edition of the State of the Internet report, we may
consider aligning our definition of broadband with the
FCC’s in future editions of the report.) In contrast to
the “high broadband” and “broadband” classifications,
the “narrowband” data included below is for connections
to Akamai slower than 256 Kbps. Note that the percent-
age changes reflected below are not additive – they
are relative to the prior quarter(s). (That is, a Q1 value
of 50% and a Q2 value of 51% would be reflected here
as a 2% change.) A quarter-over-quarter change is shown
within the tables in several sections below in an effort
to highlight general trends, and year-over-year changes
are shown to illustrate longer-term trends.
As the quantity of HD-quality media increases over time,
and the consumption of that media increases, end users
are likely to require ever-increasing amounts of bandwidth.
A connection speed of 2 Mbps is arguably sufficient for
standard definition TV-quality video content, and 5 Mbps
for standard-definition DVD quality video content, while
Blu-Ray (1080p) video content has a maximum video bit
rate of 40 Mbps, according to the Blu-Ray FAQ.5 In addition
to average connection speeds, we continue to report
average peak6 connection speeds (formerly referred to
as “average maximum” connection speeds) around the
world, from a country/region, state, and city perspective.
This metric can provide insight into the peak speeds that
users can likely expect from their Internet connections.
Finally, as was done last quarter as well, traffic from known
mobile network providers will be analyzed and reviewed
in a separate section of the report; mobile network data
has been removed from the data set used to calculate
the metrics reported in the present section.
SECTION 4:
Geography– Global
By virtue of the hundreds of billions of requests for Web content that it services on a daily basis through its globally-deployed server network, Akamai has a unique level of visibility into the connection speeds of end-user systems and, consequently, of broadband adoption around the globe. Because Akamai has implemented a distributed network model, deploying servers within edge networks, it can deliver content more reliably and more consistently at those speeds, in contrast to centralized competitors that rely on fewer deployments in large data centers. For more information on why this is possible, please see Akamai’s How Will The Internet Scale? White Paper3 or the video explanation at www.akamai.com/whytheedge.
10 © 2010 Akamai Technologies, Inc. All Rights Reserved
4.1 Global Average Connection Speeds Globally, average connection speeds appeared to
rebound from the first quarter decline, increasing on
both a quarterly and a yearly basis. In addition, as shown
in Figure 5, South Korea posted a significant growth
in the second quarter, on both a quarterly and a yearly
basis as well. (Removal of additional mobile network
traffic from South Korea’s data set may have influenced
these figures.) Among the remainder of the top 10, five
other countries/regions saw their average connection
speeds decline in second quarter, though three of them
exhibited positive long-term trends, with strong year-
over-year growth percentages. The quarterly declines
seen across those five countries/regions, along with the
United States, were fairly modest, and are not considered
to be out of line with short-term fluctuations seen
in the past. The ongoing observed decline in Sweden’s
average connection speed is of some concern, given
the size of the decline over the last two quarters, and
especially given the efforts to deploy high-speed
connectivity across the country.7
With traffic from known mobile networks now removed
from the data set, it is not immediately clear what is causing
these observed declines in average connection speeds. The
base data set may still include usage from mobile networks
not yet identified, or from network providers that are mixing
mobile and fixed network traffic on a single autonomous
system. Or, it may point to higher usage of network-reliant
applications and devices that are consuming some amount
of network bandwidth in communicating with non-
Akamai systems.
During the second quarter, 92 countries had average
connection speeds below 1 Mbps, four fewer than the prior
quarter. Akamai measured average connection speeds below
100 Kbps in only three countries (Cuba, Tonga, and Mayotte)
in the second quarter – down from five in the first quarter,
but consistent with the fourth quarter of 2009. (Note that
the slowest countries often have the smallest number of
unique IP addresses connecting to Akamai, so it may be the
case that a few more countries fell below that 1000 unique
IP address threshold than in the first quarter.) Mayotte con-
tinued to be the country with the lowest average connection
speed, remaining consistent at 40 Kbps.
SECTION 4:
Geography– Global (continued)
Figure 5: Average Measured Connection Speed by Country/Region
– Global 1.8 3.8% 6.1%
1 South Korea 17 38% 47%
2 Hong Kong 8.6 -4.9% 23%
3 Japan 8.0 2.1% 9.9%
4 Romania 6.8 7.4% 8.4%
5 Netherlands 6.5 8.5% 22%
6 Latvia 6.3 -1.0% 24%
7 Sweden 5.5 -11% -7.7%
8 Czech Republic 5.3 -2.9% -0.7%
9 Belgium 5.3 10% 11%
10 Denmark 5.2 -2.8% 11%
…
16 United States 4.6 -1.8% 1.8%
YoY ChangeQoQ ChangeQ2 ‘10 Avg. MbpsCountry/Region
16
7
6
4
3
2
1
8
9
5
10
11© 2010 Akamai Technologies, Inc. All Rights Reserved
4.2 Global Average Connection Speeds, City View As we have done in the past three editions of the
State of the Internet report, we again examine average
measured connection speeds at a city level. As we did
in the first quarter, we have applied “filters” for unique
IP address count (50,000 or more seen by Akamai in the
second quarter) and academic institutions (removing data
from known academic networks). In addition, as with
the other data sets used in Section 4 of this report, traffic
from known mobile networks has been removed as well.
Once again, Masan, South Korea held the title as the
city with the fastest average connection speed. Masan’s
average speed increased by more than 5 Mbps quarter-
over-quarter, making it the only city in the world with an
average connection speed higher than 20 Mbps. Twenty
other cities, however, achieved average connection speeds
in excess of 10 Mbps. These cities were mostly in South
Korea and Japan, though Europe’s fastest city (Umea,
Sweden) was in that group as well. Yokohama, Japan
fell just short, missing the 10 Mbps threshold by just
under 1 Kbps. Umea, Sweden, had an average connection
speed of 10.3 Mbps, up from 9.8 Mbps in the first quar-
ter. In North America, the fastest city was Victoria, British
Columbia at 7.2 Mbps, down from 7.8 Mbps in the first
quarter. In the United States, Monterey Park, CA remained
the fastest city, with an average connection speed of
6.9 Mbps, down from 7.2 Mbps in the first quarter.
Cities in Asia continued to dominate the top 100 list,
accounting for three quarters of the top 100, as shown
in Figure 6. This included 62 cities in Japan, and 12 cities
in South Korea, and Hong Kong. Europe accounted for
15 cities across eight countries, with Romania ranking
highest with five cities in the top 100. North America
accounted for the final ten, with two in Canada and
the remaining eight in the United States. The span of
average connection speeds grew in the second quarter
as well, from 9.5 Mbps separating the first and hundredth
city in the first quarter to just over 14.7 Mbps separating
Masan, South Korea and Riga, Latvia in the second quarter.
Cities in Asia dominate the Global Average Connection Speeds top 100 list, with 62 located in Japan, 12 more in South Korea, and Hong Kong. Ten cities from North America made it into the top 100, along with 15 cities in Europe.
12 © 2010 Akamai Technologies, Inc. All Rights Reserved
Figure 6: Average Measured Connection Speed, Top Global Cities
1 South Korea Masan 20.9
2 South Korea Taegu 19.5
3 South Korea Poryong 18.9
4 South Korea Seocho 18.0
5 South Korea Kimchon 17.7
6 South Korea Milyang 17.3
7 South Korea Ilsan 17.1
8 South Korea Seoul 16.0
9 South Korea Anyang 15.3
10 South Korea Suwon 14.9
11 Japan Shimotsuma 12.2
12 Japan Tokai 12.2
13 Japan Kanagawa 12.2
14 Japan Usen 11.8
15 South Korea Inchon 11.4
16 Japan Asahi 11.1
17 Japan Urawa 11.0
18 Japan Tochigi 10.8
19 Japan Hiroshima 10.7
20 Sweden Umea 10.3
21 Japan Shizuoka 10.1
22 Japan Yokohama 10.0
23 Japan Ibaraki 10.0
24 Japan Kyoto 9.5
25 Japan Nagoya 9.4
26 Japan Gifu 9.2
27 Japan Marunouchi 9.2
28 Japan Kobe 9.2
29 Japan Hyogo 9.1
30 Japan Nara 8.9
31 Japan Wakayama 8.8
32 Japan Chiba 8.8
33 Japan Nagano 8.7
34 Japan Fukuoka 8.7
35 Japan Hakodate 8.6
36 Japan Yokkaichi 8.6
37 Japan Sendai 8.6
38 Japan Niho 8.5
39 Romania Constanta 8.5
40 South Korea Yongsan 8.5
41 Japan Otsu 8.4
42 Japan Hodogaya 8.4
43 Japan Matsuyama 8.2
44 Japan Hamamatsu 8.2
45 Switzerland Geneva 8.2
46 Japan Soka 8.1
47 Japan Tokushima 8.1
48 Japan Kanazawa 8.1
49 Japan Kagawa 8.1
50 Japan Fukui 8.1
51 Hong Kong Hong Kong 8.0
52 Japan Tokyo 7.9
53 Japan Niigata 7.9
54 Romania Iasi 7.9
55 Japan Mito 7.9
56 Japan Yamaguchi 7.9
57 Japan Kochi 7.8
58 Netherlands Groningen 7.8
59 Japan Kokuryo 7.8
60 Japan Utsunomiya 7.7
61 Germany Baden-Baden 7.6
62 Japan Yamagata 7.6
63 Japan Okayama 7.5
64 Japan Kumamoto 7.5
65 Norway Lyse 7.5
66 Japan Kagoshima 7.4
67 Romania Timisoara 7.4
68 Japan Yosida 7.4
69 Japan Toyama 7.3
70 Japan Osaka 7.3
71 Japan Kofu 7.2
72 Canada Victoria, BC 7.2
73 Japan Miyazaki 7.2
74 Japan Iwaki 6.9
75 Japan Tottori 6.9
76 United States Monterey Park, CA 6.9
77 United States Riverside, CA 6.8
78 Netherlands Joure 6.8
79 Japan Saga 6.7
80 Netherlands Tilburg 6.7
81 Japan Oita 6.6
82 Japan Nagasaki 6.6
83 Portugal Coimbra 6.5
83 Japan Naha 6.5
85 Canada Kelowna, BC 6.5
86 Japan Otemachi 6.5
87 Japan Morioka 6.5
88 Japan Okidate 6.5
89 Japan Akita 6.4
90 United States Union, NJ 6.4
91 United States Oakland, CA 6.4
92 Romania Clujnapoca 6.3
93 Romania Bucharest 6.3
94 United States Fairfield, CA 6.3
95 United States Walnut Creek, CA 6.2
96 United States Boston Metro, MA 6.2
97 United States Hayward, CA 6.2
98 Japan Sapporo 6.1
99 Netherlands Kamperland 6.1
100 Latvia Riga 6.1
Q2 ‘10 Avg. Mbps Q2 ‘10 Avg. MbpsCity CityCountry/Region Country/Region
SECTION 4:
Geography– Global (continued)
13© 2010 Akamai Technologies, Inc. All Rights Reserved
4.3 Global Average Peak Connection SpeedsThe average peak connection speed metric represents
an average of the maximum measured connection
speeds across all of the unique IP addresses seen by
Akamai from a particular geography. The average is
used in order to mitigate the impact of unrepresentative
maximum measured connection speeds. In contrast
to the average measured connection speed, the average
peak connection speed metric is more representative
of what many end-user Internet connections are capable
of. (This includes the application of so-called speed
boosting technologies that may be implemented within
the network by providers, in order to deliver faster
download speeds for some larger files.) Note that
data from known mobile networks has also been
removed from the source data set for this metric.
As shown in Figure 7, average peak connection speeds
demonstrated quarterly growth on a global basis, in nine
of the top 10 countries/regions, and in the United States.
(The lone exception is Sweden, whose declining average
connection speed was also discussed above in Section
4.1.) While the quarterly increases were generally pretty
modest, South Korea, Belgium and Monaco turned
in impressive double-digit percentage growth. Extremely
impressive growth was also seen in the Maldives, which
saw its average peak connection speed nearly triple
quarter-over-quarter. However, the Maldives only had
approximately 12,000 unique IPs connecting to Akamai in
the second quarter, so that must be considered in compari-
son to other countries/regions in the top 10, most of whom
had millions of unique IP addresses connecting to Akamai.
Once again, Asia led this metric, with South Korea, Hong
Kong, and Japan taking the top three slots, with average
peak connection speeds in excess of 30 Mbps observed
in South Korea and Hong Kong. European countries took
six of the seven remaining top 10 slots, with the United
States falling to 11th place (from 8th place in the first
quarter). Average peak connection speeds increased year-
over-year by at least double digit percentages globally, as
well as across the top 10 countries and the United States.
In looking at the average peak speed distribution around
the world, only South Korea and Hong Kong exceeded
30 Mbps – it appears that the aggressive pricing for
100 Mbps and 1 Gbps connections in Hong Kong8 may
be seeing commercial success. Four more countries/regions
registered average peak connection speeds in excess of 20
Mbps, while 49 others had average peak connection speeds
in excess of 10 Mbps. In looking at the “high broadband”
threshold of 5 Mbps, 55 additional countries exceeded that.
Only six countries saw average peak connection speeds
below 1 Mbps, and Mayotte ceded its last-place mantle
to Tonga, whose 365 Kbps average peak connection speed
was just over 5x its average speed for the second quarter.
Figure 7: Average Peak Connection Speed by Country/Region
– Global 6.8 6.4% 23%
1 South Korea 38 16% 24%
2 Hong Kong 32 7.2% 31%
3 Japan 28 8.1% 11%
4 Romania 27 7.7% 31%
5 Latvia 20 5.3% 57%
6 Maldives 20 287% 790%
7 Belgium 19 15% 49%
8 Sweden 18 -3.7% 15%
9 Monaco 18 25% 59%
10 Portugal 17 5.9% 31%
…
11 United States 16 1.2% 18%
YoY ChangeQoQ ChangeQ2 ‘10 Peak MbpsCountry/Region
11
8
5
4
3
2
1
6
7
10
9
14 © 2010 Akamai Technologies, Inc. All Rights Reserved
4.4 Global Average Peak Connection Speeds, City ViewAs we first did in the 1st Quarter, 2010 State of
the Internet report, we again examine average peak
connection speeds at a city level, applying “filters” for
unique IP address count (50,000 or more seen by Akamai
in the second quarter) and academic institutions (removing
data from known academic networks). In addition, as with
the other data sets used in Section 4 of this report, traffic
from known mobile networks has been removed.
In looking at the distribution of average peak speeds seen
across the top 100 cities, we found that more than two-
thirds of the top 100 cities fell within the 20-30 Mbps range,
as shown in the histogram in Figure 8. The largest individual
cluster was between 23-24 Mbps, with 16 cities having
average peak connection speeds in that band.
With an average peak connection speed of just over
44 Mbps, up approximately 10% from last quarter,
Masan, South Korea once again topped the list,
as shown in Figure 9. At a city level, it appears that average
peak connection speeds generally increased during the
second quarter, with eight cities now registering average
peak connection speeds above 40 Mbps, up from just three
cities in the first quarter. Among the balance of the top
100 cities, 32 additional cities have average peak connection
speeds above 30 Mbps, and the remaining 60 cities were
all above 20 Mbps. The speed of the last city in the top
100 was up just about 5%, or approximately 1 Mbps,
quarter-over-quarter as well.
Cities in Asia dominated this metric as well, with the
top 100 list including Hong Kong, 57 cities from Japan
and 12 from South Korea. North America charted 21 cities
among the top 100, all from the United States. Monterey
Park, CA was the United States city with the highest average
peak connection speed, measured at nearly 26 Mbps. In
Europe, nine cities from six countries rounded out the top
100, with Constanta, Romania the European city with the
highest average peak connection speed, at nearly 38 Mbps.
SECTION 4:
Geography– Global (continued)
Figure 8: Distribution of Average Peak Speeds by Mbps Band
Num
ber o
f Citi
es
Mbps
18
16
14
12
10
8
6
4
2
023 29 3925 3531 4127 3733 43 4422 28 3824 3430 4026 3632 42
15© 2010 Akamai Technologies, Inc. All Rights Reserved
Figure 9: Average Peak Connection Speed, Top Global Cities
1 South Korea Masan 44.4
2 South Korea Taegu 44.0
3 Japan Tokai 42.3
4 Japan Kanagawa 42.2
5 South Korea Poryong 42.0
6 Japan Marunouchi 41.8
7 South Korea Kimchon 40.9
8 South Korea Seocho 40.3
9 Japan Shimotsuma 39.7
10 South Korea Milyang 38.2
11 South Korea Ilsan 38.2
12 Japan Urawa 38.1
13 Romania Constanta 37.7
14 Japan Hiroshima 36.9
15 South Korea Suwon 36.4
16 South Korea Seoul 36.3
17 Japan Yokohama 36.3
18 Japan Hodogaya 35.7
19 Japan Tochigi 35.4
20 Japan Usen 35.4
21 Japan Asahi 35.3
22 South Korea Anyang 34.1
23 Romania Iasi 33.8
24 Japan Soka 33.8
25 Japan Kobe 33.2
26 Japan Nagano 33.2
27 Japan Kyoto 32.6
28 Japan Chiba 32.6
29 Japan Fukuoka 32.3
30 Japan Nagoya 32.2
31 Japan Ibaraki 32.2
32 Romania Timisoara 31.6
33 Japan Gifu 31.5
34 Japan Shizuoka 30.7
35 Japan Yokkaichi 30.7
36 Hong Kong Hongkong 30.4
37 Japan Niho 30.4
38 Japan Wakayama 30.3
39 Japan Sendai 30.3
40 Japan Kokuryo 30.0
41 Japan Nara 29.9
42 Japan Niigata 29.8
43 Japan Mito 29.8
44 Japan Fukui 29.5
45 Japan Kanazawa 29.3
46 Japan Utsunomiya 29.2
47 Japan Otsu 29.2
48 Japan Yosida 28.9
49 Japan Hamamatsu 28.9
50 Japan Hakodate 28.8
51 Japan Kagawa 28.2
52 Japan Tokushima 27.7
53 Japan Yamaguchi 27.5
54 Germany Baden-Baden 27.3
55 Japan Matsuyama 27.1
56 South Korea Inchon 27.0
57 Japan Iwaki 26.9
58 Portugal Coimbra 26.8
59 Japan Tokyo 26.7
60 Japan Kofu 26.5
61 Japan Toyama 26.4
62 Japan Kochi 26.3
63 South Korea Yongsan 26.1
64 Japan Osaka 26.1
65 Japan Yamagata 25.9
66 United States Monterey Park, CA 25.9
67 United States Everett, WA 25.7
68 United States Olympia, WA 25.7
69 United States Bellevue, WA 25.6
70 United States Federal Way, WA 25.6
71 Japan Okayama 25.6
72 Sweden Umea 25.3
73 Romania Bucharest 25.3
74 Norway Lyse 25.2
75 Japan Okidate 25.1
76 Japan Hyogo 24.8
77 Japan Morioka 24.7
78 Japan Kumamoto 24.4
79 Japan Sapporo 24.2
80 United States Hickory, NC 24.0
81 United States Oakland, CA 24.0
82 Japan Kagoshima 23.8
83 United States Fairfield, CA 23.8
84 United States Hayward, CA 23.6
85 United States Salem, OR 23.5
86 United States Traverse City, MI 23.5
87 United States Vancouver, WA 23.4
88 United States Beaverton, OR 23.4
89 Japan Akita 23.3
90 United States Tacoma, WA 23.3
91 United States Riverside, CA 23.3
92 United States Boston Metro, MA 23.3
93 United States Fond Du Lac, WI 23.2
94 United States San Jose, CA 23.1
95 Switzerland Geneva 23.1
96 United States Walnut Creek, CA 23.0
97 Japan Naha 22.9
98 United States Aurora, CO 22.8
99 Japan Miyazaki 22.8
100 United States Arvada, CO 22.7
Q2 ‘10 Peak Mbps Q2 ‘10 Peak MbpsCity CityCountry/Region Country/Region
16 © 2010 Akamai Technologies, Inc. All Rights Reserved
SECTION 4:
Geography– Global (continued)
4.5 Global High Broadband ConnectivityIn the second quarter of 2010, global high broadband
adoption continued to grow, reaching 22%. As shown
in Figure 10, South Korea again was the country with
the highest level of high broadband adoption, with
impressive double-digit quarterly growth to 75%,
pushing it well ahead of second-place Japan, at 60%.
Among the top 10 countries, strong double-digit levels
of quarterly growth were also observed in the Nether-
lands, Belgium, and Moldova. Consistent with the
observations in Sections 4.1 and 4.3, Sweden saw
a significant decline in high broadband adoption in
the second quarter, and was the only country in the
top 10 to show a yearly decline as well. The United
States experienced a slight decline in the second
quarter, though the yearly trend is still positive.
Looking at high broadband adoption on a global
basis, South Korea and Japan are the only two countries
where more than half of the connections to Akamai
are at speeds greater than 5 Mbps. Beyond that, there
are 13 additional countries/regions where more than
a quarter of the connections are at high broadband
rates and 22 more where one in ten connections
to Akamai are faster than 5 Mbps. Europe is best
represented among this aggregate set of countries, with
31 countries having greater than 10% high broadband
adoption – only seven countries are included from the Asia
Pacific region. The overall global trends for high broadband
adoption are positive as well, with 1.6x as many countries
showing quarterly gains as losses, and 1.7x as many
countries showing yearly gains as losses.
Data on the speed distribution of high broadband
connectivity had been presented in its own sub-section
in prior reports, but will be folded into the larger high
broadband connectivity discussion starting with this
edition of the State of the Internet report. In examining
the speed distributions on a global basis, as well as for
the top 10 countries, it appears that the greatest quarterly
shifts occurred, in most cases, among the highest speed
‘buckets’ (20-25 Mbps and 25+ Mbps). For example,
in the second quarter, nearly 21% of South Korea’s
connections to Akamai were at speeds above 25 Mbps
– nearly double the level seen in the prior quarter. Double
digit percentage increases for this bucket were also seen
for the global figure (27% growth), as well as in the
Netherlands (12% growth), Romania (25% growth), and
Belgium (21% growth). The United States, unfortunately,
Figure 10: High Broadband Connectivity, Fastest Countries/Regions
– Global 22% 0.7% 6.5%
1 South Korea 75% 12% 8.3%
2 Japan 60% 1.1% 7.6%
3 Netherlands 49% 13% 34%
4 Romania 48% 0.1% 7.5%
5 Hong Kong 46% -1.0% 19%
6 Latvia 43% 4.5% 75%
7 Belgium 40% 11% 23%
8 Denmark 37% -8.5% 18%
9 Moldova 34% 27% 101%
10 Sweden 34% -21% -21%
…
13 United States 30% -5.1% 3.8%
YoY ChangeQoQ Change% Above 5 MbpsCountry/Region
13
10
6
9
42
5
1
3
7
8
17© 2010 Akamai Technologies, Inc. All Rights Reserved
appears to have lost some ground here in the second
quarter, with less than 1% of connections to Akamai at
speeds over 25 Mbps, a decline of nearly 10% from the
prior quarter. While the United States posted a long-term
year-over-year decline of 15% for this metric, the picture
is more positive globally and for other countries in the
top 10. Yearly growth in connections to Akamai at speeds
over 25 Mbps was seen in the global figure (35% growth),
as well as in eight of the top 10 countries, ranging from
an anemic 0.8% increase in Sweden to an astonishing
increase of over 700% in Moldova. (To be fair, the pool
of unique IP addresses from Moldova connecting to
Akamai at high broadband speeds was much smaller
than other countries in the top 10, so large percentage
increases must be considered with that in mind.)
We expect that, on a global basis, as the adoption and
rollout of DOCSIS 3.0 technology by cable Internet provid-
ers, as well as other FTTH initiatives by telecom providers,
become more widespread the percentage of connections
in these higher speed ‘buckets’ will continue to grow over
time. Ideally, the growth observed in the 25+ Mbps bucket
in the second quarter across the top 10 countries, as well
as a number of other countries around the world, is the
beginning of a long-term trend, and not simply a one-
quarter event.
4.6 Global Broadband ConnectivityAs shown in Figure 11, the quarter-over-quarter changes
for global broadband connectivity were fairly minor in the
second quarter, with Monaco seeing the largest increase,
at just under four percent. Broadband adoption rates in
Hong Kong and Belgium remained constant from the first
quarter, while the Isle of Man and Slovakia had variances
of just one-tenth of a percent. (It is worth noting that both
Monaco & Isle of Man recorded less than 20,000 unique
IPs for this metric, so while their broadband adoption
levels are certainly impressive, they need to be considered
in that context, as other countries among the top 10 have
hundreds of thousands or millions of unique IP addresses
connecting to Akamai at speeds in excess of 2 Mbps.)
With the exception of an extremely small decline in South
Korea, broadband adoption increased year over year at
a global level, for nine of the top 10 countries, and in
the United States.
The levels of broadband adoption continue to trend ever
higher overall as well, with eight countries/regions seeing
adoption levels of 90% or better, and nearly three-quarters
of the connections to Akamai from the United States
at speeds of 2 Mbps or above. In the second quarter of
2010, 55 countries/regions had broadband adoption levels
in excess of 50% -- up from 50 countries/regions in the
first quarter, and 41 countries/regions in the same period
a year earlier.
Figure 11: Broadband Connectivity, Fast Countries/Regions
– Global 59% 3.3% 4.1%
1 Monaco 95% 3.8% 8.5%
2 South Korea 93% 2.2% -0.1%
3 Isle Of Man 93% 0.1% 5.3%
4 Hong Kong 92% – 0.7%
5 Switzerland 91% -0.2% 0.3%
6 Belgium 91% – 0.4%
7 Bulgaria 91% 2.2% 14%
8 Latvia 90% 1.9% 19%
9 Czech Republic 88% 3.2% 5.2%
10 Slovakia 88% -0.1% 2.9%
…
37 United States 72% 1.2% 4.8%
YoY ChangeQoQ Change% Above 2 MbpsCountry/Region
37
8
7
4
2
5
6
3
10
9
1
18 © 2010 Akamai Technologies, Inc. All Rights Reserved
4.7 Global Narrowband ConnectivityIn line with the general trends for increased broadband
and high broadband adoption that have been discussed
in prior sections, it appears that levels of narrowband
adoption dropped in many countries around the world
during the second quarter. As shown in Figure 12, Cuba
was the only country among the top 10 that recorded a
quarterly increase, and an almost insignificant one at that.
(However, this 95% narrowband adoption rate causes
the relatively few Internet users in the country to “suf-
fer through agonizingly long waits to open an e-mail, let
alone view a photo or video,” according to the United
Nations International Telecommunications Union.9)
In the second quarter, less than five percent of connections
to Akamai overall were at speeds below 256 Kbps. This was
the case in 65 countries around the world, with South Korea
having the lowest levels of narrowband adoption, at 0.2%.
In the United States, narrowband adoption dropped 10%
in the second quarter to 2.8%, a level 33% lower than the
same period a year ago. Globally, over 130 countries had
lower levels of narrowband adoption in the second quarter,
compared with over 35 that had higher levels.
SECTION 4:
Geography– Global (continued)
There is no broadband in Cuba and the relatively few Internet users in the country suffer through agonizingly long waits to open an e-mail, let alone view a photo or video. This also hampers government and business operations. [http://www.reuters.com/article/idUSTRE65H3H320100618]
Figure 12: Narrowband Connectivity, Slowest Countries/Regions
– Global 4.6% -13% -12%
1 Mayotte 98% -0.5% -1.1%
2 Wallis And Futuna 96% -2.1% -3.2%
3 Cuba 95% 0.1% 0.7%
4 Madagascar 90% -1.8% -3.2%
5 Vanuatu 89% -4.7% -7.2%
6 Guyana 87% -6.1% -1.2%
7 Equatorial Guinea 86% -12% -12%
8 Niger 86% -2.7% 1.5%
9 Rwanda 83% -9.4% -11%
10 Uzbekistan 82% -3.4% 2.6%
…
119 United States 2.8% -10% -33%
YoY Change
QoQ Change
% Below 256 Kbps
Country/Region
118 7
3
1
42
5
6
108
9
19© 2010 Akamai Technologies, Inc. All Rights Reserved
5.1 United States Average Connection SpeedsThe overall average connection speed for the United
States as a whole in the second quarter of 2010 was
4.6 Mbps. As in the first quarter, this was exceeded by
22 states, including those in the top 10, as shown in
Figure 13. Both across the country and within the top
10, most quarterly changes were relatively minor – the
biggest growth was seen in West Virginia and Rhode
Island, gaining 11% and 10% respectively, while
the biggest losses were seen in Iowa and Arizona,
declining 11% and 10% respectively. Year-over-year
SECTION 5:
Geography– United States
The metrics for the United States presented here are based on a subset of the data used for Section 4, and are subject to the same thresholds and filters discussed within the prior section. (The subset used for this section includes connections identified as coming from networks in the United States, based on classification by Akamai’s EdgeScape10 geolocation tool.)
changes were more significant, with 11 states posting
double digit increases, led by Alaska’s yearly growth of
44%, and 34 states in total getting faster since the same
quarter a year ago. In fact, Alaska’s second quarter growth
helped it move out of being the state with the slowest
average connection speed, ceding that dubious honor this
quarter to Iowa, which had an average connection speed
of 2.8 Mbps. (Ahead of Iowa by only a mere 96 Kbps,
it is certainly possible that Alaska could fall back to the
last-place slot again in the future.)
Figure 13: Average Measured Connection Speed by State
1 Delaware 7.2 -4.5% 15%
2 Rhode Island 6.2 10% 16%
3 New Hampshire 5.9 1.3% -6.4%
4 District Of Columbia 5.6 -5.7% -1.1%
5 Massachusetts 5.5 -7.1% -3.2%
6 Nevada 5.5 2.4% 5.8%
7 Maryland 5.3 -5.2% -7.0%
8 Vermont 5.3 -2.9% -4.7%
9 California 5.2 -2.3% 8.3%
10 Utah 5.2 -6.4% 2.1%
YoY Change
QoQ Change
Q2 ‘10 Avg. Mbps
State
4
1
25
8 3
7
10 9
6
20 © 2010 Akamai Technologies, Inc. All Rights Reserved
5.2 United States Average Connection Speeds, City View As with the Global Average Connection Speeds, City
View presented in Section 4.2, connections from known
academic networks were removed from the underlying
data set for this metric, and the 50,000 unique IP
address filter was used for this view as well.
In reviewing the top 10 cities in the United States
with the highest average connection speeds, as shown
in Figure 14, we see that cities in California once again
dominate the list, holding seven of the top 10 slots.
However, we also note that the average connection
speeds for these cities in California all suffered quarterly
declines, dropping 300-400 Kbps on average. Interesting-
ly, the other three cities in the top 10 list were different
than those included in the first quarter list – cities
in Maryland, New York, and Washington were replaced
by cities in New Jersey, Massachusetts, and Illinois.
SECTION 5:
Geography– United States (continued)
Figure 14: Average Measured Connection Speed, Top United States Cities by Speed
1 Monterey Park, CA 6.9
2 Riverside, CA 6.8
3 Union, NJ 6.4
4 Oakland, CA 6.4
5 Fairfield, CA 6.3
6 Walnut Creek, CA 6.2
7 Boston Metro, MA 6.2
8 Hayward, CA 6.2
9 Mount Prospect, IL 6.1
10 San Mateo, CA 6.1
Q2 ‘10 Avg. MbpsCity
5.3 United States Average Peak Connection SpeedsThe overall average peak connection speed calculated
by Akamai for the United States as a whole for the second
quarter of 2010 was 16 Mbps. This was exceeded by 18
states and the District of Columbia, including all of those
in the top 10, as shown in Figure 15. For this metric, Idaho
once again ranked lowest, with an average peak connection
speed of 8.9 Mbps, down nearly 5% from the first quarter.
In addition, Idaho was the only state to record an average
peak connection speed below 10 Mbps in the second
quarter – in the first quarter, Wyoming and Arkansas
had done so as well.
Quarterly trending of average peak connection speeds
was comparatively modest, with the greatest change seen
in West Virginia, which had a quarterly loss of 11%. Overall,
28 states and the District of Columbia saw higher average
peak connection speeds in the second quarter – in contrast,
48 states and the District of Columbia saw growth in the
first quarter. The quarterly increases were more muted as
well, as no state saw a quarterly increase greater than 10%,
whereas seven states exceeded this level in the first quarter.
While quarterly declines were seen in 22 states, in most of
them, these declines amounted to a few hundred Kbps or
less, so there does not appear to be any reason for concern
here, as these variances are within reasonable ranges. The
year-over year picture was much rosier, with no observed
yearly declines. Forty-five states and the District of Columbia
saw yearly growth of 10% or more, with 21 of those states
growing 20% or more. Montana was the big gainer for this
metric as well, increasing its average peak connection speed
by 45% since the second quarter of 2009. Alaska was a
close second for yearly growth, increasing its average peak
connection speed by 43% year-over-year.
21© 2010 Akamai Technologies, Inc. All Rights Reserved
In looking at the complete list of qualifying cities, we once
again note that there were no cities from Delaware, North
Dakota, or Vermont on the list. Additionally, we observed
that the average peak connection speeds in all of the listed
cities exceeded 2 Mbps, though the majority of the cities
had average peak connection speeds between 10 and
20 Mbps.
Figure 16: Average Peak Connection Speed, Top United States Cities by Speed
1 Monterey Park, CA 25.9
2 Everett, WA 25.7
3 Olympia, WA 25.7
4 Bellevue, WA 25.6
5 Federal Way, WA 25.6
6 Hickory, NC 24.0
7 Oakland, CA 24.0
8 Fairfield, CA 23.8
9 Hayward, CA 23.6
10 Salem, OR 23.5
Q2 ‘10 Avg. MbpsCity
5.4 United States Average Peak Connection Speeds, City ViewOnce again, the city view of average peak connection
speeds in the United States was dominated by cities on
the West Coast, as illustrated in Figure 16. California
saw four cities in the top 10, as did Washington, and
one from Oregon made the list as well. Hickory, NC
was the lone East Coast city in the top 10. In contrast
to the general quarterly declines seen in the top 10
United States cities for the average connection speed
metric, it appears that average peak connection speeds
among the top 10 cities generally improved from the
first quarter – Monterey Park was ~600 Kbps faster
than a quarter ago, and last quarter’s 10th place city,
Fairfield, CA, moved up to 8th place this quarter,
adding approximately 300 Kbps.
Figure 15: Average Peak Connection Speed by State
1 Delaware 24 -5.4% 20%
2 Rhode Island 21 0.1% 23%
3 New Hampshire 21 1.1% 5.3%
4 Hawaii 21 2.8% 10%
5 District Of Columbia 20 0.9% 13%
6 Massachusetts 19 -3.1% 14%
7 California 19 2.3% 27%
8 Washington 19 4.7% 16%
9 Maryland 18 -0.6% 16%
10 Vermont 18 -6.3% 9.0%
YoY Change
QoQ Change
Q2 ‘10 Peak Mbps
State
5
1
26
10 3
97
8
4
22 © 2010 Akamai Technologies, Inc. All Rights Reserved
SECTION 5:
Geography– United States (continued)
5.5 United States High Broadband ConnectivityIn the second quarter of 2010, nine of the top 10 states
saw lower levels of high broadband adoption than in the
first quarter, as shown in Figure 17. The lone standout
was Rhode Island, which increased high broadband adop-
tion by more than a third quarter-over-quarter. Across
the whole country, 37 states and the District of Columbia
saw high broadband adoption levels decline, with North
Dakota, Iowa, and Arizona each losing 20% or more.
Of the 12 states that posted quarterly gains, four states
saw growth in excess of 10%, and Arkansas joined Rhode
Island with a quarterly growth rate of more than 30%.
In the first quarter of 2010, we posited that the declines
observed may have been due, at least in part, to the
delivery of streaming video for the 2010 Winter Olympics
over the Akamai HD Network. We believe that increased
second quarter consumption of long form video, such
as movies and television shows, at bitrates below 5 Mbps,
may have contributed to the widespread quarterly declines
discussed above. As was discussed in the 1st Quarter,
2010 State of the Internet report, Akamai plans to
implement filtering of such rate-limited content, as
appropriate, from future State of the Internet data sets.
Year-over-year changes were slightly more positive, with
27 states and the District of Columbia increasing their levels
of high broadband adoption. Alaska led the pack here,
growing high broadband adoption by 233% over the prior
year, while South Dakota achieved a very respectable 93%
yearly growth rate. (Both states had posted the highest yearly
growth rates in the first quarter as well.) Twenty-one states
(including South Dakota and Alaska) increased more than
10% year-over-year. Of the 23 states that lost ground since
the second quarter of 2009,14 had declines of 6% or less,
while five declined 20% or more.
Similar to section 4.5 above, data on the speed distribution
of high broadband connectivity had been presented in its
own sub-section in prior reports, but will be folded into the
larger high broadband connectivity discussion starting with
this edition of the State of the Internet report. In looking
at a state-level breakdown of speeds above 5 Mbps, it is
no surprise that the largest percentage of connections falls
within the 5-10 Mbps bucket, accounting for anywhere from
6.3% to 55% of connections to Akamai. At the high end,
connections to Akamai in excess of 25 Mbps account for
anywhere from 0.1% to just 3.2% of connections
Figure 17: High Broadband Connectivity, Fastest U.S. States
1 Delaware 67% -6.6% 45%
2 New Hampshire 53% -2.6% -5.9%
3 Rhode Island 52% 37% 47%
4 New Jersey 45% -4.8% -5.3%
5 Maryland 44% -5.8% -5.0%
6 District Of Columbia 44% -0.7% 5.3%
7 Massachusetts 43% -8.5% -0.5%
8 Nevada 40% -3.3% 8.2%
9 Connecticut 37% -10% -3.3%
10 New York 37% -13% -22%
YoY Change
QoQ Change
% Above 5 Mbps
State
6
14
39
710
2
5
8
23© 2010 Akamai Technologies, Inc. All Rights Reserved
to Akamai. While four states grew their percentage
of 25+ Mbps connections to Akamai in excess of 100%
in the second quarter, it is important to note that even
with that growth rate, their total percentage of 25+
Mbps connections each remained under 1% Of the
seven total states that saw yearly growth in the percent-
age of 25+ Mbps connections in the second quarter,
only two of them exceeded 1%.
It is clear that there is still a long way to go in bringing
high speed connectivity to a significant portion of the
American populace. Of course, one challenge in doing
so is the associated cost. A report11 issued by the United
States Federal Communications Commission in May
noted that to reach the nation’s estimated 14 million
un-served citizens with broadband service of at least 4
Mbps, it would cost $23.5 billion to deliver using DSL
and fixed wireless, and $62 billion to deploy fiber to the
home. Note that this estimate is based on targeting
only a portion of the national population, and only
at minimum speeds of 4 Mbps – enabling higher speed
connectivity for more people would carry a significantly
higher price tag.
5.6 United States Broadband ConnectivityBroadband adoption within the United States continues
to be strong, with Delaware gradually closing in on the
elusive 100% mark, and six other states with broadband
adoption of 80% or more, as shown in Figure 18. Across
the country, 47 states and the District of Columbia had more
than half of their connections to Akamai at speeds in excess
of 2 Mbps during the second quarter. The remaining three
states (Wyoming, Idaho, and Iowa) were not that far behind,
with adoption rates of 49%, 44%, and 41% respectively.
In contrast to the quarterly changes discussed above in
Section 5.5, quarterly growth was seen in eight of the top
10 states, and 32 states plus the District of Columbia overall.
The observed growth was fairly modest, with Delaware
inching up just 0.2%, to Mississippi’s 8.5% quarter-over-
quarter increase. Year-over-year numbers were very positive,
with only four states experiencing declining broadband
adoption since the second quarter of 2009 – Nebraska,
Maine, North Dakota, and Vermont. Montana saw
the biggest yearly increase, at 56%, bringing it to a
broadband adoption level of 55% in the second quarter.
Figure 18: Broadband Connectivity, Fast U.S. States
1 Delaware 98% 0.2% 1.5%
2 New Hampshire 90% 0.5% 2.3%
3 Rhode Island 89% 4.3% 5.3%
4 Hawaii 87% 0.2% 3.1%
5 Connecticut 84% 1.0% 2.6%
6 Nevada 82% 1.7% 5.6%
7 Maryland 80% 1.6% 2.9%
8 Maine 79% -0.4% -3.2%
9 New York 79% -0.1% 0.2%
10 New Jersey 78% 1.1% 3.6%
YoY Change
QoQ Change
% Above 2 Mbps
State
1
35
892
4
10
7
6
24 © 2010 Akamai Technologies, Inc. All Rights Reserved
SECTION 5:
Geography– United States (continued)
It is interesting to note the contrasting quarterly trends
between the high broadband and broadband adoption
metrics in the United States for the second quarter. As
noted above, one possible explanation for the observed
quarterly trends of declining high broadband adoption
and increased broadband adoption is significant growth
in the consumption of long-form media content (movies,
television shows) at bitrates between 2-5 Mbps, delivered
by Akamai using HTTP streaming techniques.
5.7 United States Narrowband Connectivity The percentage of connections to Akamai at speeds
below 256 Kbps from U.S. states continues a general
descent, highlighting a trend over time to higher speed
connectivity. As shown in Figure 19, nine of the top
10 states with the highest percentages of narrowband
(<256 Kbps) connections saw quarter-over-quarter
declines, with only North Dakota seeing an increase, and
a rather small one at that. Across the whole United States,
42 states and the District of Columbia saw narrowband
adoption drop from first quarter levels – Mississippi saw
the greatest decline, dropping 46%. Year-over-year trends
improved greatly as compared to those observed in the first
quarter, with only West Virginia seeing a higher narrowband
adoption rate as compared to the second quarter of 2009.
(And with a narrowband adoption rate of only 2.5%,
it is still doing comparatively well.)
In the second quarter, four states (Hawaii, Rhode Island,
Nevada, and Delaware) had narrowband adoption rates
of less than 1%. Delaware was the lowest at 0.2%, which
is not surprising, given that it also has the highest adoption
rates for high broadband and broadband connectivity, as well
as the highest average and average peak connection speeds.
Figure 19: Narrowband Connectivity, Slowest U.S. States
1 Alaska 7.7% -4.0% -22%
2 District Of Columbia 6.5% -7.3% -32%
3 Missouri 5.7% -6.7% -30%
4 Iowa 5.5% -4.0% -25%
5 Georgia 4.5% -8.8% -36%
6 Illinois 4.0% -5.1% -28%
7 North Dakota 3.9% 0.7% -37%
8 Ohio 3.8% -10% -30%
9 Colorado 3.8% -8.2% -31%
10 Texas 3.7% -8.4% -32%
YoY Change
QoQ Change
% Below 256 Kbps
State
2
1
10 5
3
4
7
9 6 8
25© 2010 Akamai Technologies, Inc. All Rights Reserved
SECTION 6:
Mobile
Building on the data presented in the State of the Internet reports over the last several quarters, Akamai continues to identify additional mobile networks for inclusion in the report. As we did in the 1st Quarter, 2010 State of the Internet report, we are including information on the average connection speed and average peak connection speed by provider, as well as insight into the average megabytes (MB) downloaded from Akamai per month per unique IP address associated with mobile networks. In addition, we have moved the discussion of attack traffic originating from mobile networks into this section. As was noted last quarter, the source data set for this section is subject to the following constraints:
• A minimum of 1,000 unique IP addresses connecting
to Akamai from the network in the second quarter
of 2010 was required for inclusion in the list.
• In countries where Akamai had data for multiple
network providers, only the top three are listed,
based on unique IP count.
• The names of specific mobile network providers have
been anonymized, and providers will be identified
by a unique ID.
• Data is included only for networks where Akamai
believes that the entire Autonomous System (AS) is
mobile – that is, if a network provider mixes traffic from
fixed/wireline (DSL, cable, etc.) connections with traffic
from mobile connections on a single network identifier,
that AS was not included in the source data set.
• Akamai’s EdgeScape database was used for the
continental assignments.
6.1 Attack Traffic, Top Originating Countries When the set of attack traffic observed by Akamai during
the second quarter of 2010 is limited to that seen from
known mobile network providers, the top 10 list of attack
sources is rather different in a number of ways. For one,
half of the top 10 countries are different – Chile, the United
Kingdom, Malaysia, Poland, and Ireland appear in Figure 20,
but not in the overall list in Figure 1 (although Poland has
been known to be a top 10 attack traffic source at times).
Another difference is the concentration – over 40% of the
observed mobile attack traffic comes from just the top two
countries (as opposed to the top five overall), and the top
10 countries are responsible for nearly 80% of the observed
attacks, in contrast to nearly 60% overall.
Figure 20: Attack Traffic from Mobile Networks, Top Originating Countries/Regions
1 Italy 25%
2 Brazil 18%
3 Chile 7.5%
4 United Kingdom 6.2%
5 Malaysia 5.0%
6 Poland 4.8%
7 China 3.2%
8 Russia 2.9%
9 United States 2.7%
10 Ireland 2.4%
– Other 22%
% TrafficCountry/Region
9
3
6
5
8
2
71
4
10
26 © 2010 Akamai Technologies, Inc. All Rights Reserved
6.2 Attack Traffic, Top Ports Not surprisingly, the top 10 ports targeted by attack
traffic coming from mobile networks were fairly similar
to overall attack traffic observed by Akamai. As shown
in Figure 21, nine of the top 10 ports targeted by mobile
attack traffic are the same as in Figure 3, which highlights
the ports targeted by all attack traffic observed by Akamai.
The lone standout here is Port 2967 (Symantec System
Center), instead of Port 3389 (Microsoft Terminal Services),
at the bottom of the top 10 list. Attack traffic coming
from mobile networks also appears to be more
SECTION 6:
Mobile (continued)
concentrated than the overall set of attack traffic, with
Port 445 accounting for 84% of mobile attacks (vs. only
62% overall), and the top 10 ports accounting for more
than 98% of observed mobile attacks (vs. just over 86%
overall). As noted previously, we believe that the observed
attack traffic that is originating from known mobile
networks is likely being generated by infected PC-type
clients connecting to wireless networks through mobile
broadband technologies, and not by infected smartphones
or similar mobile devices.
Other 5.6%
Figure 21: Attack Traffic from Mobile Networks, Top Ports
445 Microsoft-DS 84%
23 Telnet 4.9%
22 SSH 3.3%
135 Microsoft-RPC 3.1%
139 NetBIOS 0.8%
4899 Remote Administrator 0.7%
1433 Microsoft SQL Server 0.6%
5900 VNC Server 0.4%
80 WWW 0.3%
2967 Symantec System Center 0.1%
Various Other 1.8%
% TrafficPort UsePort WWW 0.3%Symantec System Center 0.1%
Other 1.8%Microsoft-RPC 3.1%
SSH 3.3%
Remote Administrator 0.7%
Microsoft SQL Server 0.6%NetBIOS 0.8%
VNC Server 0.4%
Microsoft-DS 84%
Telnet4.9%
27© 2010 Akamai Technologies, Inc. All Rights Reserved
SECTION 6:
Mobile (continued)
6.3 Mobile Connection Speeds & Data ConsumptionIn examining the data shown in Figure 22, we see that
there is once again an extremely wide range of average
connection speeds, and curiously, the fastest (6.1 Mbps)
and slowest (115 Kbps) were once again seen on mobile
providers in Slovakia. Of the 109 mobile providers listed,
19 had average connection speeds in the broadband
(2 Mbps or above) range (up from 14 in the first quarter),
while 29 more had average measured connection speeds
of 1 Mbps or more (up from 21 in the first quarter). On
a quarter-over-quarter basis, mobile connectivity appeared
to improve in many places, with users on 95 of the 109
listed mobile providers experiencing higher connection
speeds in the second quarter. These improvements ranged
from a 0.9% increase on a provider in Singapore to
doubled speeds on a provider in Saudi Arabia. Quarterly
losses were fairly nominal, with the worst a 31% average
connection speed decline on a Nigerian mobile provider.
Year-over-year changes were also very positive, ranging
from connection speeds that more than quadrupled on
the aforementioned Saudi Arabian provider, to the slight
0.3% growth seen from a provider in the Czech Republic.
There were three providers across Japan, France, and Hong
Kong that showed no change year-over-year – the Hong
Kong provider also showed no quarterly change, which
is unusual.
In reviewing the average peak connection speed data,
we observed that the Slovakian provider with the highest
average connection speed topped this metric as well,
with an average peak connection speed of just over
20 Mbps. Note that this excludes the United Kingdom
mobile provider listed as having an average peak con-
nection speed in excess of 36 Mbps – we believe that
this provider uses a mobile gateway/proxy architecture
that reflects gateway/proxy-to-Akamai communication
speeds, rather than mobile device-to-Akamai communication
speeds. Akamai continues to investigate methods of mitigat-
ing the impact of these gateways/proxies on the source data
sets that will be used for future editions of the State of the
Internet report. Average peak connection speeds proved to
be very strong in the second quarter, with all but six of the
listed providers exceeding 1 Mbps. Trends over time were
very positive as well, with only 11 providers seeing quarterly
declines in average peak connection speeds, and only 11 see-
ing yearly declines as well. Interestingly, those same providers
in Japan, Hong Kong, and France remained flat quarter-over-
quarter and year-over-year for this metric as well.
As we did in the first quarter, we are again examining the
average amount (MB) of data downloaded from Akamai
per month per unique IP address seen from that mobile pro-
vider. Not including two providers where we believe proxy/
gateway architectures are in use, we found five providers
in Canada, Puerto Rico, Slovakia, Germany, and Austria
whose users, on average, consumed more than one gigabyte
(1 GB) of content from Akamai per month during the second
quarter. An additional 80 mobile providers around the world
had more than 100 MB of data downloaded from Akamai
per unique IP address per month during the second quarter
of 2010. The two providers whose users downloaded the
least data from Akamai during the second quarter were
located in Belgium, consuming 16 MB and 14 MB per unique
IP address per month respectively. In contrast to the speed
metrics discussed above, 71 mobile providers saw measured
consumption grow on a quarterly basis, while it declined on
37 providers. Year-over-year changes were more aggressive,
however, with 85 providers increasing the average amount
(MB) of data downloaded from Akamai per month per
unique IP address as compared to the second quarter
of 2009.
28 © 2010 Akamai Technologies, Inc. All Rights Reserved
SECTION 6:
Mobile (continued)
Figure 22: Average and Average Peak Connection Speed, Average Megabytes Downloaded per Month by Mobile Provider
AFRICA
Egypt EG-1 420 2172 149
Morocco MA-1 657 5420 422
Nigeria NG-1 167 3478 237
South Africa ZA-1 522 959 216
China CN-1 1972 5138 128
Hong Kong HK-1 2534 11721 849
Hong Kong HK-2 2154 8892 229
Hong Kong HK-3 1929 4802 19
Indonesia ID-1 269 7730 14879
Israel IL-1 1107 5328 82
Japan JP-1 1546 6145 113
Kuwait KW-1 1933 10306 306
Malaysia MY-1 354 2004 147
Malaysia MY-2 2286 n/a 565
Malaysia MY-3 729 4307 461
Pakistan PK-1 815 4684 394
Saudi Arabia SA-1 1927 7003 222
Singapore SG-2 654 6359 101
Singapore SG-3 1265 5046 299
South Korea KR-1 1482 2957 34
Sri Lanka LK-1 630 3348 183
Taiwan TW-1 988 4073 104
Taiwan TW-2 501 2650 133
Thailand TH-1 733 6514 105
Austria AT-1 2745 10427 133
Austria AT-2 2208 7598 1119
Belgium BE-1 2435 8353 303
Belgium BE-2 954 2358 16
Belgium BE-3 298 749 14
Croatia HR-1 1050 4137 45
Czech Republic CZ-1 1066 4268 78
Czech Republic CZ-2 462 2292 118
Czech Republic CZ-3 2604 8416 140
Estonia EE-1 825 3626 197
France FR-1 489 2424 140
France FR-2 1742 5391 609
France FR-3 579 3821 200
Germany DE-1 486 2258 61
Germany DE-2 2447 8503 1246
Greece GR-1 1207 7851 512
Greece GR-2 575 4129 119
Hungary HU-1 1423 6798 145
Hungary HU-2 1862 7887 106
Ireland IE-1 2104 8897 357
Ireland IE-2 1302 8020 357
Ireland IE-3 1129 8409 514
Italy IT-1 1131 7144 358
Italy IT-2 2007 7343 286
Italy IT-3 2739 10292 423
Italy I T-4 594 4453 248
Lithuania LT-1 1616 7987 280
Lithuania LT-2 944 4083 205
Moldova MD-1 1100 4073 72
Moldova MD-2 1418 5528 120
Netherlands NL-1 917 2067 20
Q2 ‘10 Avg. Kbps
Q2 ‘10 Avg. Kbps
ID IDQ2 ‘10 Peak Kbps
Q2 ‘10 Peak Kbps
Q2 ‘10 Avg. MB/month
Q2 ‘10 Avg. MB/month
Country/Region Country/Region
ASIA
EURoPE
AFRICA
It is clear that as digital media consumption grows; mobile will be the front and center of this evolution. [http://www.chetansharma.com/blog/2009/06/03/silicon-india-article-mobile-media-evolution]
29© 2010 Akamai Technologies, Inc. All Rights Reserved
Netherlands NL-2 1698 3347 19
Norway NO-1 977 3850 56
Norway NO-2 1287 4382 65
Poland PL-1 3675 10583 111
Poland PL-2 1005 3869 40
Poland PL-3 682 3581 121
Portugal PT-1 376 1514 32
Romania RO-1 471 2467 79
Russia RU-1 5287 18979 129
Russia RU-2 773 2813 57
Russia RU-3 571 2378 104
Slovakia SK-1 115 726 33
Slovakia SK-2 2166 6485 1703
Slovakia SK-3 6129 20203 625
Slovenia SI-1 1274 5114 48
Spain ES-1 1358 7951 274
Spain ES-3 891 5032 168
Ukraine UA-1 441 1303 34
United Kingdom UK-1 1285 7529 386
United Kingdom UK-2 2115 9446 526
United Kingdom UK-3 3650 36549 12787
Canada CA-1 2788 13338 7363
Canada CA-2 826 2166 499
El Salvador SV-1 536 3465 203
El Salvador SV-2 880 6047 470
El Salvador SV-3 763 4097 721
Guatemala GT-1 392 2655 119
Guatemala GT-2 496 4148 474
Mexico MX-2 767 5837 474
Mexico MX-3 459 4119 354
Netherlands Antilles AN-1 403 2544 192
Nicaragua NI-1 704 4680 279
Puerto Rico PR-1 2306 9021 2326
United States US-1 967 2282 28
United States US-2 980 2838 27
United States US-3 910 3011 351
Australia AU-1 974 7741 819
Australia AU-3 1457 6199 137
Guam GU-1 327 852 177
New Caledonia NC-1 656 2526 268
New Zealand NZ-2 1098 6018 303
Argentina AR-1 358 3206 116
Argentina AR-2 463 3425 165
Bolivia BO-1 156 2269 166
Brazil BR-1 600 3104 108
Brazil BR-2 501 4338 126
Chile CL-3 612 5453 382
Chile CL-4 730 4939 446
Colombia CO-1 570 5249 193
Paraguay PY-1 297 2209 113
Paraguay PY-2 312 2844 278
Uruguay UY-1 727 5357 245
Venezuela VE-1 575 3668 367
Q2 ‘10 Avg. Kbps
Q2 ‘10 Avg. Kbps
ID IDQ2 ‘10 Peak Kbps
Q2 ‘10 Peak Kbps
Q2 ‘10 Avg. MB/month
Q2 ‘10 Avg. MB/month
Country/Region Country/Region
NoRTh AMERICA
SoUTh AMERICA
oCEANIA
Kids of the now generation are growing with connected electronics that is fundamentally altering the behaviors and expectations of interaction, communication, consumption, and monetization. [http://www.chetansharma.com/usmarketupdateq22010.htm]
30 © 2010 Akamai Technologies, Inc. All Rights Reserved
Unique IP Addresses
Avg. Connection Speed (Mbps)
Peak Connection Speed (Mbps)
% Attack Traffic
% Above 5 Mbps
% Above 2 Mbps
% Below 256 Kbps
Country/Region
SECTION 7:
Appendix
EURoPE
Austria 0.2% 2,296,675 3.8 11.4 18% 65% 1.0%
Belgium 0.1% 3,513,996 5.3 19.0 40% 91% 0.3%
Czech Republic 0.2% 1,674,706 5.3 14.5 34% 88% 0.3%
Denmark 0.2% 2,160,122 5.2 14.0 37% 87% 0.7%
Finland 0.1% 2,338,594 4.1 11.5 23% 56% 1.3%
France 1.8% 22,740,013 3.4 11.8 12% 73% 0.8%
Germany 2.9% 31,220,592 4.1 13.8 20% 85% 0.8%
Greece 0.3% 2,124,500 3.0 12.8 5.2% 71% 1.5%
Iceland 0.0% 121,363 4.3 14.7 16% 81% --
Ireland 0.1% 1,307,681 5.1 13.7 15% 65% 3.1%
Italy 3.5% 10,845,562 3.0 11.5 6.0% 76% 3.1%
Luxembourg 0.0% 160,771 3.8 12.2 14% 82% 1.2%
Netherlands 0.4% 6,983,247 6.5 16.0 49% 88% 0.6%
Norway 0.4% 2,377,412 4.7 13.7 24% 80% 0.8%
Portugal 0.8% 2,263,510 3.9 17.0 24% 78% 0.3%
Spain 1.8% 11,444,687 2.7 10.7 5.8% 63% 1.0%
Sweden 0.2% 4,406,626 5.5 18.5 34% 73% 1.6%
Switzerland 0.2% 2,629,299 5.0 15.4 24% 91% 0.6%
United Kingdom 1.2% 20,042,782 3.9 13.4 17% 83% 1.0%
ASIA/PACIFIC
Australia 0.4% 8,842,180 2.8 11.0 12% 50% 4.9%
China 0.5% 60,042,886 0.9 3.4 0.3% 6.6% 13%
Hong Kong 0.4% 2,242,583 8.6 31.7 46% 92% 0.5%
India 1.9% 4,634,437 0.8 4.7 0.5% 4.3% 30%
Japan 2.6% 34,833,550 8.0 27.9 60% 86% 1.4%
Malaysia 1.0% 1,530,039 1.2 6.7 0.8% 6.0% 6.6%
New Zealand 0.3% 1,347,976 3.3 12.7 12% 71% 6.3%
Singapore 0.3% 1,602,979 3.1 12.8 17% 55% 4.9%
South Korea 1.5% 16,795,947 16.6 38.0 75% 93% 0.2%
Taiwan 6.0% 6,251,406 4.1 14.0 23% 66% 1.1%
MIDDlE EAST
Egypt 1.6% 1,077,113 0.8 5.3 0.5% 6.4% 9.4%
Israel 0.7% 1,960,429 3.3 11.2 6.0% 76% 0.3%
Kuwait 0.2% 245,545 1.3 7.5 1.4% 17% 7.1%
Saudi Arabia 0.5% 1,495,923 1.8 6.6 0.5% 33% 1.0%
Sudan 0.0% 25,900 0.5 3.2 -- -- 21%
Syria 0.0% 90,045 2.0 3.6 7.6% 45% 28%
United Arab Emirates (UAE) 0.2% 763,650 1.5 7.4 4.3% 18% 7.2%
lATIN & SoUTh AMERICA
Argentina 2.3% 3,933,382 1.5 6.8 0.7% 23% 3.4%
Brazil 5.7% 12,003,674 1.4 5.6 2.3% 19% 16%
Chile 0.5% 2,019,604 2.2 10.1 2.6% 44% 1.4%
Colombia 1.0% 2,408,231 1.6 7.2 0.4% 27% 2.8%
Mexico 0.4% 8,001,433 1.5 6.4 0.5% 16% 1.8%
Peru 1.5% 664,123 1.2 6.5 0.7% 8.3% 2.6%
Venezuela 0.3% 2,183,054 0.7 3.8 -- 1.6% 15%
NoRTh AMERICA
Canada 1.7% 11,644,413 4.7 15.1 31% 84% 1.9%
United States 11% 131,397,342 4.6 16.4 30% 72% 2.8%
31© 2010 Akamai Technologies, Inc. All Rights Reserved
1 The http://isc.sans.edu/diary.html?storyid=8881
2 http://isc.sans.edu/diary.html?storyid=9031
3 http://www.akamai.com/dl/whitepapers/How_will_the_internet_scale.pdf
4 http://www.fcc.gov/Daily_Releases/Daily_Business/2010/db0720/FCC-10-129A1.pdf
5 http://www.blu-ray.com/faq/
6 The “average peak connection speed” metric represents an average of the peak measured connection
speeds across all of the unique IP addresses seen by Akamai from a particular geography. The average
is used in order to mitigate the impact of unrepresentative peak measured connection speeds. In contrast
to the average measured connection speed, the average peak connection speed metric is more representative
of what many end-user Internet connections are capable of. (This includes the application of so-called speed
boosting technologies that may be implemented within the network by providers, in order to deliver faster
download speeds for some larger files.)
7 http://www.telegraph.co.uk/technology/broadband/6500162/Swedish-government-pledges-super-fast-
broadband-for-all.html
8 http://arstechnica.com/tech-policy/news/2010/04/1gbps-symmetric-fiber-us26-in-hong-kong.ars
9 http://www.reuters.com/article/idUSTRE65H3H320100618
10 http://www.akamai.com/html/technology/products/edgescape.html
11 http://download.broadband.gov/plan/the-broadband-availability-gap-obi-technical-paper-no-1.pdf
SECTION 8:
Endnotes
The Akamai Difference Akamai® provides market-leading, cloud-based services for optimizing Web and mobile
content and applications, online HD video, and secure e-commerce. Combining highly-
distributed, energy-efficient computing with intelligent software, Akamai’s global platform
is transforming the cloud into a more viable place to inform, entertain, advertise, transact
and collaborate. To learn how the world’s leading enterprises are optimizing their business
in the cloud, please visit www.akamai.com and follow @Akamai on Twitter.
AcknowledgementsEDITOR: David Belson
CONTRIBUTOR: Jon Thompson
CONTRIBUTOR: Patrick Gilmore
CONTRIBUTOR:Alloysius Gideon
EXECUTIVE EDITOR: Brad Rinklin
EXECUTIVE EDITOR:Tom Leighton
Please send comments, questions, and corrections to [email protected]
Follow @akamai and @akamai_soti on
Akamai | Powering A Better Internet™
For more information, visit www.akamai.com
©2010 Akamai Technologies, Inc. All Rights Reserved. Reproduction
in whole or in part in any form or medium without express written
permission is prohibited. Akamai and the Akamai wave logo
are registered trademarks of Akamai Technologies, Inc. Other
trademarks used herein may be owned by other companies and
are used for descriptive purposes only. Akamai believes that the
information in this publication is accurate as of its publication date;
such information is subject to change without notice.
International Offices
Unterfoehring, GermanyParis, FranceMilan, ItalyLondon, EnglandMadrid, SpainStockholm, Sweden
Bangalore, IndiaSydney, AustraliaBeijing, ChinaTokyo, JapanSeoul, KoreaSingapore
Akamai Technologies, Inc.
U.S. Headquarters
8 Cambridge CenterCambridge, MA 02142Tel 617.444.3000Fax 617.444.3001U.S. toll-free 877.4AKAMAI(877.425.2624)
www.akamai.com